From MWCSWiki

---

What is OKM?

<b>OKM</b> is a multifaceted research project designed to explore various aspects of supporting "subjectivity" in collaborative knowledge bases. It is currently composed of the following related subprojects:

• OKM interface - The design of an intuitive and easy-to-use collaborative knowledge base interface, so that even novices to the Semantic Web can effectively author RDF data in a distributed setting.
• Human interaction studies - A set of usability studies to determine what novices find difficult about authoring Semantic Web knowledge, and how an interface like OKM could be designed to counteract those difficulties.
• Divergent data - A particular aspect of OKM is that it permits multiple conflicting views of the data to simultaneously co-exist in the knowledge base, and to be coherently browsed.
• Category-based browsing - An experimental interface that allows users to group objects according to their own subjective perceptions, and then employs data mining techniques to learn from those groupings so that the user can ask questions about the larger knowledge base as a whole.
• Linked Data Integration - Extending OKM to traverse the web of linked data, subsume it on demand into our KB, and allow the user to make assertions about external URIs.

ESWC 2010 paper

ESWC2010poster

Revised KR 2010 paper

TO DO LIST BEFORE MONDAY

Who is OKM?

OKM is currently staffed by:

• Stephen Davies
• Jesse Hatfield
• Chris Donaher
• Stacey Aylor
• Jessica Zeitz
• Trillane Burlar

Other contributors in the past have included Shane Voisard and Jonathan Morin.

Where is OKM?

What should everybody be working on right now? Here's a shot at answering that question:

• Stephen
  1. Clean up database and launch cinefile experiment
  2. Lit search for studies on "obstacles users encounter when creating formal knowledge representations (like RDF)"
  3. Lit search "reification"/"n-ary relationships" to see if there's any "how hard do people find this" stuff
  4. Finish "grading" experiment 1 when it comes in (w/ Jessica, Jesse, Chris)
     • Reification paragraphs (5, 6, 7)
     • First-half paragraphs (finish Jason)
  5. Outline for KR paper
  6. Design experiment #2 (w/ Jessica, Jesse, Chris, Trillane)
• Jesse
  1. play around with the system and determine a good threshold for keyword significance
  2. play around with the system and determine whether presence-only and presence-absence are both worth testing (or whether we should ditch one of them)
  3. Finishing the end-to-end test code for the cinefile experiment, including storing test data
  4. Peruse the "Tagging" book to see if there's anything relevant to the way we're using keywords
  5. "Grade" experiment 1 when it comes in (w/ Jessica, Stephen, Chris)
• Chris
  1. Implement other changes to OKM interface (w/ Trillane)
  2. "Grade" experiment 1 when it comes in (w/ Jessica, Stephen, Jesse)
  3. Begin to implement import procedure (later in fall)
• Stacey
  1. <strike> Research JavaScript toolkits (including Prototype and Ext/JS) and figure out which one(s) can do the kind of icons-on-a-canvas-connected-by-lines stuff that we need to do </strike>
  2. Download (http://www.imdb.com/interfaces#plain) and import (from IMDB) other movie attributes that we can have the user operate on during analysis. I'm thinking some nominal ones (studio, producer, writer, director) and numeric ones (year, gross earnings, production cost).
  3. <strike> Download and install Firebug. Also, look at jslint.org: it is an extremely useful tool in javascript debugging. </strike>
• Jessica
  1. <strike>Write email to KR conference</strike>
  2. Phelps rubric
  3. Research concept maps (Joseph Novak)
  4. <strike>Conduct experiment 1</strike>
  5. "Grade" experiment 1 when it comes in (w/ Stephen, Jesse, Chris)
  6. Actually play around with other RDF editing tools (RDFAuthor, Tabulator, Protege, OntoWiki, Kiwi) to get a feel for how they work
  7. Read up on usability testing principles. Do some big picture, strategic thinking about "how to evaluate OKM and determine how effectively people are able to use its key design features"
  8. Research YAGO
• Trillane
  1. Implement other changes to OKM interface (w/ Stephen)